Artificial Intelligence is frequently presented either as a disruptive technological force destined to replace human decision-making or as an autonomous actor capable of assuming increasingly significant governance functions. This new Working Paper argues that both interpretations fundamentally misunderstand the constitutional role of Artificial Intelligence within democratic institutions. Building upon the theoretical foundations established throughout the Institutional Cognition Research Programme, the paper develops the General Theory of Human–AI Institutional Intelligence, proposing that Artificial Intelligence should be understood neither as a replacement for human institutions nor as an independent governing authority, but as a complementary cognitive architecture operating within constitutional democratic systems.
Within the paper we first distinguish human cognition from computational cognition, demonstrating that each possesses fundamentally different yet mutually reinforcing capabilities. Human institutions retain sovereignty, constitutional legitimacy, ethical reasoning, contextual interpretation and political responsibility, while Artificial Intelligence contributes computational cognition through large-scale analysis, recursive learning, anticipatory modelling, knowledge integration and systemic pattern recognition. Rather than competing for institutional authority, these complementary forms of intelligence recursively interact within hybrid cognitive architectures that substantially expand institutional reasoning while preserving democratic governance.
In the study we subsequently develop a comprehensive theoretical framework describing constitutional Human–AI collaboration through AI-augmented institutions, hybrid institutional reasoning, recursive learning, distributed Human–AI governance, institutional co-evolution and Human–AI knowledge ecosystems. These concepts culminate in a unified theoretical architecture explaining how future institutions progressively evolve into hybrid cognitive ecosystems capable of continuously learning, adapting and governing increasingly complex societies.
The central contribution of the Working Paper is the formulation of the General Theory of Human–AI Institutional Intelligence, which establishes that the future evolution of governance depends upon preserving the constitutional distinction between computational cognition and democratic authority while enabling increasingly sophisticated collaboration between the two. Artificial Intelligence becomes institutional cognitive infrastructure rather than political agency, strengthening institutional intelligence without altering the democratic principle that sovereignty, legitimacy and public responsibility remain exclusively human.
In the paper we conclude arguing that Human–AI Institutional Intelligence represents a new stage in the historical evolution of constitutional governance, providing democratic institutions with the cognitive capacities necessary to govern twenty-first-century complexity while remaining fully aligned with constitutional democracy, human oversight and public accountability.
